browsertime: cold page load, Chrome 73 vs GVE 68 aarch 64 vs Fenix 68 aarch64 vs Fennec 64 on the Pixel 2

Per-device, per-site vehicle comparisons

The following graph gives insight into how the vehicles compare on the sites in the test corpus.

Here are the median values:

Here is the noise in the data captured. What stands out are the sheer number of outliers:

However, if we look at individual vehicles, we see that each vehicle has a similar number of outliers. That is, it’s not only one vehicle is less reliable than the others.

Gecko engine vehicles over Chrome 73